from fastapi import FastAPI, UploadFile, File
import requests
from transformers import pipeline
from sentence_transformers import CrossEncoder
import pandas as pd
import os

app = FastAPI()

ai_detector = pipeline("text-classification", model="roberta-base-openai-detector")
text_model = CrossEncoder("cross-encoder/stsb-roberta-large")

TEXT_DB = "text_plagiarism.csv"

def load_database():
    if not os.path.exists(TEXT_DB):
        pd.DataFrame(columns=["content", "plagiarism_score"]).to_csv(TEXT_DB, index=False)
    return pd.read_csv(TEXT_DB)

def save_to_database(content, plagiarism_score):
    df = load_database()
    new_entry = pd.DataFrame({"content": [content], "plagiarism_score": [plagiarism_score]})
    df = pd.concat([df, new_entry], ignore_index=True)
    df.to_csv(TEXT_DB, index=False)

@app.post("/check_text")
def check_text(text: str):
    stored_texts = load_database()["content"].tolist()
    if stored_texts:
        similarity_scores = text_model.predict([[text, stored] for stored in stored_texts])
        highest_similarity = max(similarity_scores)
    else:
        highest_similarity = 0
    save_to_database(text, highest_similarity * 100)
    return {"plagiarism_score": highest_similarity * 100}

@app.post("/detect_ai")
def detect_ai(text: str):
    result = ai_detector(text)
    return result[0]